Rock Mechanics and Rock Engineering (2020) 53:4015–4029

https://doi.org/10.1007/s00603-020-02151-x

ORIGINAL PAPER

Experimental Investigation of Gas Transfer Properties and Stress

Coupling Effects of Salt Rocks
Dongmei Zhang1,2 · Frédéric Skoczylas1,2 · Franck Agostini2 · Laurent Jeannin3

Received: 9 December 2019 / Accepted: 14 May 2020 / Published online: 27 May 2020
© Springer-Verlag GmbH Austria, part of Springer Nature 2020

Abstract
This experimental study is dedicated to the measurements of gas permeability under confinement and to preliminary char-
acterizations of salt rock poromechanical behavior. Salt porosity and permeability under loading are recorded either with
argon or hydrogen gas. Permeability measurements carried out with both gases led to similar values. Poromechanical effects
occur but they are weak and strongly dependent on the level of confining pressure. Furthermore, triaxial tests have been
used to investigate the evolution of permeability of damaged salt samples with time and confinement. These experimental
characterizations are finally analysed in the context of gas storage applications.

Keywords Salt rock · Gas permeability · Coupling effects · Poromechanical behaviours · Porosity · Time effects

1 Introduction
From an engineering perspective, salt rocks are usually
considered to be an impermeable medium and they have
revealed to be a suitable geological environment for under-
ground natural gas storage and waste disposal (Schulze et al.
2001; Bérest and Brouard 2003). In addition to their ultra-
low permeability, the ability to stand high pressure and their
self-healing capacity after damage (Chen et al. 2013a, b;
Hou 2003) make very attractive the underground storage in
salt formations (Stolten and Grube 2010). Hence caverns
have been developed for natural gas storage over several
decades.
Rock salt is a polycrystalline material made of grains of
halite, with grain size in the range 2–20 mm (for most hal-
ite rocks), containing also impurities in solid solution and
secondary mineral phases (Urai et al. 2008). The objective
of this paper is to investigate gas transfer through this kind
of rocks and to evaluate whether coupling effects may occur
* Frédéric Skoczylas
frederic.skoczylas@centralelille.fr
1
School of Civil Engineering and Architecture, Changzhou
Institute of Technology, Changzhou, China
2
University of Lille, CNRS, Centrale Lille, LaMcube
UMR9013, Lille, France
3
Storengy, 12 rue Raoul Nordling, Djinn, CS 70001,
92274 Bois Colombes Cedex, France
between pore fluid pressure and rock skeleton. Two applica-
tions deserve indeed our attention for salt permeability char-
acterization in this study. First, due to salt creeping, some
caverns lost their economic value and have to be abandoned.
Abandonment experiments of salt caverns (Karimi-Jafari
2007) demonstrate the existence of permeation at the cavern
scale. The cavern is filled with brine and sealed for abandon-
ment; an equilibrium in brine pressure can be reached when
salt mass creep and cavern shrinkage are balanced by brine
permeation through the cavern wall. A second application, in
the current context of the energy transition, can be found in
hydrogen storage (Maurice et al. 2015). Hydrogen, produced
for example by solar energy, may be stored in salt caverns.
Under this scenario, permeability related to hydrogen needs
to be assessed.
In-situ tightness well tests (Berest et al. 2001), abandon-
ment tests (Bérest et al. 2001; Karimi-Jafari 2007) and also
laboratory experiments (Gloyna and Reynolds 1961; Stor-
mont and Daemen 1992; Popp and Kern 1998; Liu et al.
2015) at the core scale have shown that salt rock has often
a low permeability. This permeability is strongly influenced
on one hand by the stress state. Several authors suggest that
permeation, observed during in situ tests in salt caverns,
had been partially induced by cavern construction itself, as
high deviatoric stresses were developed in a disturb rock
zone (denoted DRZ) around the cavern (Stormont 1997;
Hou 2003). In the DRZ, damage and dilatancy occur (Alkan
2009; Labaune et al. 2018). Permeability indeed increases

13
Vol.:(0123456789)
4016
by several orders of magnitude under deviatoric loading
inducing damage (Peach et al. 1987; Pusch et al. 2002; Xing
et al. 2014; Stormont and Daemen 1992). Permeability is
also dependant of the salt initial state as measurements on
intact, initially damaged, and compacted crushed salts may
be highly different (Brodsky 1993; Stormont 1990). On the
other hand, permeability is strongly influenced by time-
dependent effects: from a mechanical point of view, rock salt
is an elasto-viscoplastic rock, which shows a strong creeping
behavior. Its instantaneous response is elastic, while its long
term-mechanical response under constant deviatoric load-
ings is mainly time-dependent, and explains, therefore, the
volume loss and the progressive closure of the salt caverns
during gas storage operation. Understanding the variation in
permeability when rock creeps is, therefore, crucial (Chan
et al. 2001). However, there are few previous researches on
the long-term evolution of permeability (Peach 1991; Stor-
mont and Daemen 1992).
In addition, as this rock is a little permeable, internal fluid
pressure may contribute to loading and be involved in the
stress–strain relationship characterizing the equation of state
of the material and may influence permeability at a given
loading. This phenomenon is often observed on more or less
permeable other sedimentary rocks, but, to our knowledge,
has little been studied for salt.
To perform both hydraulic and poromechanical measure-
ments, a specific experimental set-up has been developed and
is presented in the first section of this paper. The transient
pulse decay method under loading was carried out for per-
meability measurements. The use of gas (argon in this study)
for poromechanical coupling experiments allowed to detect
tenuous coupling effects compared with the use of water
or brine. Hydrostatic compression tests were performed to
investigate isotropic stress and fluid pressure influences on
permeability and porosity. Comparisons between hydrogen
and argon permeability show that argon permeability may
be considered as representative of hydrogen permeability.
In a next section, preliminary poromechanical experiments
evidenced that the rock salt skeleton is mechanically coupled
to fluid i.e. pore fluid pressure is able to induce volumetric
expansion. The concept of effective stress may, therefore,
applies, but the estimated Biot’s coefficients are far below
the usual values measured for sandstones, shales or other
sedimentary rocks (Wang et al. 2018; Yuan et al. 2017). A
high sensitivity of the permeability and Biot’s coefficient
to confinement was observed and this is generally related
to a material with cracks or grain boundaries (Chen et al.
2013a, b). Hence, the last part is dedicated to conventional
triaxial tests to investigate the evolution of permeability in
damaged salt samples. Following this operation, damaged
samples were tested under hydrostatic loadings to explore
healing effects and poromechanical coupling of salt rocks
in the vicinity of cavern walls. This long time study (more
13
D. Zhang et al.
than 18 months) has not focused on a specific property and
has been conducted to put in light some crucial aspects in
the salt behaviour, which can be very useful when hydrogen
storage or salt cavern abandonment have to be designed and
planned.
2 Experimental Material and Strategy
2.1 Material Used for the Experiments
The salt rock cores, shown in Fig. 1a, come from a cenozoic
sedimentary basin in East of France. Sample preparation was
the first tricky task as salt rocks cannot be cored with water
that dissolves the material. Air coring had been unsuccessful
(cutting tool generates particles, that agglomerate and do not
allow further cutting) and it was neither possible to core with
a hydrocarbon fluid in our laboratory.
It was finally chosen a coring with saturated brine by
modifying the water supply circuit of our corer. This was
successful and cores could be obtained close to the required
diameter of 65 mm to match with our experimental cells.
As it is shown in Fig. 1, the salts used are composed of
large crystals (centimetric size), which lead to a heteroge-
neous material. Different gauge sizes were hence tested and
this surprisingly showed that the strain results were not sig-
nificantly influenced by gauge size. For practical reasons,
the best choice revealed to be the use of crossed gauges of
20 mm length (Fig. 1g). Two sets of gauges were glued at
diametrically opposed positions on the cylindrical surface of
the specimen. Before gluing gauges a thin layer of fast-cured
epoxy was applied to the sample to smooth its rough surface
and then polished. A thin layer of silicone sealant was then
applied on the sample surface to protect the gauges. The
sample was then sealed by a rubber membrane and placed
into the cell. The experimental apparatus consisting of a
Fig. 1  Salt core and salt rock samples with gauges
Experimental Investigation of Gas Transfer Properties and Stress Coupling Effects of Salt… 4017

confining cell, which can reach confining pressures Pc as

high as 100 MPa, together with a gas injection device, is
shown in Fig. 2a. The confining pressure was controlled by
a Gilson pump. The gas used in this experiment is 99%-
pure Argon (or sometimes hydrogen) and the value of gas
pressure is continuously recorded using a LabView system.
As mentioned before, a triaxial cell was also used for the
purpose of damaging the sample, this cell is similar to the
hydrostatic one but equipped with a piston to apply the axial
Fig. 3  Schematic diagram of the transient pulse decay method
stress (Fig. 2b).

2.1.1 Gas Permeability (Pulse‑Decay Pressure Method) the flow. Walder and Nur (1986) stated that the decrease
in pressure difference ΔP obeys the following power
For extremely low permeability materials, such as this function:
rock salt, the transient pulse decay method is convenient
instead of the steady state flow method. It is based on ΔP(t) = ΔP0 e−ct , (1)
a small pressure increase ∆P0 at one end of the sample
for which pore gas pressure is initially P io (Liang et al.
( )
A×K 1 1
2001; Frédéric et al. 2003; Fedor et al. 2008; Hsieh et al.
With c = Pf + , (2)
𝜇×L V1 V2
1981). The advantage is undeniable here as Pio can also be
used as the pore fluid pressure intended to detect potential where c depends on the: permeability K, sample dimensions
coupling effects. The schematic diagram of this method L (length) and A (section surface), reservoir volumes V1 and
is shown in Fig. 3. An initial pressure difference ΔP 0 V2, gas viscosity μ (2.2 × 10–5 Pa·s) for Argon at room tem-
between the upper and lower surfaces of the sample is perature. Pf is the final equilibrium pressure, which can be
applied (at the upper one) and then slowly decreases with calculated from
V1 × ΔP
Pf = Pi0 + , (3)
V1 + V2

where Pi0 is the homogeneous initial pressure. The gas per-

meability is obtained according to relations (1), (2) and (3):
( )
c × 𝜇 × L V1 × V2
K= . (4)
A × Pf V1 + V2

This method [with the use of rel. (1), (2), (3)] is consid-
ered to be valid if ∆P is small compared to Pio, generally
less than 5% of Pio and if the sample porosity volume is
weak compared to the reservoir volumes. It will be seen
in the following that the material porosity is less than 1%,
the second condition is, therefore, easy to be fulfilled. On
an other hand, the analysis of such a test is based on the
validity of Darcy’s law. The latter is assumed despite the
very low values of the gas permeability obtained. Obvi-
ously those very low values rule out the occurrence of
turbulent flow. As it will be seen in the following, the gas
flow is likely to occur in cracks or gain joints. The level of
permeability will be, therefore, assumed to mainly depend
on crack density and not on the flow nature through the
crack. As a consequence the assumption of a laminar flow
(i.e. Darcy’s law validity) remains consistent.

Fig. 2  The experimental apparatus

13
4018
2.1.2 Gas‑Accessible Porosity
The gas-accessible porosity test is a highly useful tool
designed in our lab for studying the variations of porosity
under different confinement (or loading). This test can be
carried out in a confining (or triaxial) cell and allows the
variations in pore volume to be measured under loading or
unloading. The sample, for which pore volume is Vpore , is
connected to the experimental set-up consisting of a buffer
reservoir and tubes having a calibrated volume Vt (Fig. 4).
The test is carried out under isothermal conditions, and, over
the range of gas pressure used, the ideal gas law applied:
PV = nRT. (5)
The reservoir is set with an initial gas pressure Pini . The
valve (Fig. 4) is opened and the gas pressure is continu-
ously monitored by a LabView system until stabilization.
The pressure settles to a final value Pf such as:
(6)
( )
Pini Vr = nRT = Pf Vr + Vt + Vpore ,
where Vr is the volume of the reservoir, Vt is the volume of
different connected tubes and the dead volume of the cell.
Both are carefully calibrated before the test. Pf is the final
stable gas pressure value. According to relation (6), Vpore
can be calculated and leads to the connected “gas porosity”:
Φ = Vpore ∕Vsample .
As previously mentioned the relation (6) only applies
under isothermal conditions. This was ensured by placing
the whole device in a big climatic chamber in which the
Fig. 4  Device for the measurement of porosity
13
D. Zhang et al.
temperature was regulated at 21 °C for the whole test dura-
tion (generally several weeks).
2.1.3 Poro‑mechanical Properties Under Hydrostatic Stress
(Biot Coefficient)
Most studies of the poromechanical behavior of rocks are
based on works initiated by Biot (1941) and Terzaghi (1943)
for fully-fluid saturated porous media. Both mechanical
stress 𝝈 and internal fluid pressure Pi contribute to load-
ing. For an elastic porous medium, the effective stress 𝛔′ is
defined as follows, in which isotropic or orthotropic behavior
is assumed:
𝛔 = 𝛔� − BPi , (7)
B is known as the Biot’s tensor of the material. A com-
mon expression for soil is Terzaghi’s effective stress tensor
𝛔� = 𝛔 + 1Pi. For materials with a compressible solid phase,
Biot had generalized Terzaghi’s concept leading to the use
of the 2nd tensor B.
The first state equation of poroelasticity can be written
in the form:
𝛔 = ℂd ∶ 𝛆 − BPi , (8)
where 𝛆 is the strain tensor, ℂd is the drained stiffness tensor
of the porous material. For a unit change of pore (respec-
tively, confining) pressure, two strain tensors are obtained:
𝜺1 Pc = 1 = −ℂ−1 ∶ 𝜹;𝜺2 Pi = 1 = ℂ−1
( ) ( )
d d
∶ B.
It is, therefore, obvious that pore pressure coupling effects
take place if 𝜺2 is not null for an increase in pore pressure.
For an isotropic material the Biot tensor B = b1, “b” is
known as the Biot’s coefficient and can be estimated from
two tests: hydrostatic loading and pore pressure variation,
which give, respectively, the drained bulk modulus Kb and
the H modulus such as:
ΔPc = −Kb Dev , (9)
in which ∆Pc is a variation in confining pressure and ∆εv
the resultant volumetric strain.
ΔPi = HDev , (10)
in which ∆Pi is a variation in pore pressure and ∆εv the
resultant volumetric strain.
The Biot’s coefficient “b” can then be calculated by the
ratio Kb/H.
In its initial version, the Biot’s theory can be applied
for an elastic medium only. It is why, the Biot’s coefficient
measurements, based on the use of Kb and H, need unload-
ing step for ∆Pc and loading step for ∆Pi. Both of these
steps lead to a decrease in effective confining pressure,
Experimental Investigation of Gas Transfer Properties and Stress Coupling Effects of Salt… 4019

which is likely supposed to result in elastic strains. The
behavior of the rock salt is unlikely to be elastic and its
isotropy is still to be evidenced. This second assumption
is far below the scope of this study. It is, however, neces-
sary to process to unloading in Pc or to pore pressure Pi
increases to benefit from some elastic behaviour of the
material. On another hand, it must be pointed out that,
as mentioned before, coupling effects have almost never
been really investigated in salt (to our knowledge). The
purpose of this work is to evidence that, in some cases,
fluid-skeleton couplings can occur, which is the most
important point. As a consequence, the Biot’s coefficient
calculations, proposed in the following, will be mainly
done to answer single questions: is there coupling or not?
Is it “strong” or not? Hence the calculations will be based
on the evaluation of the ratio − ∆εv(∆Pi = ∆Pc)/∆εv(∆Pc),
which is equivalent to the ratio Kb/H for isotropic material.
On another hand this ratio would lead to an “apparent”
Biot’s coefficient, which can mainly be used to highlight
the fluid-skeleton coupling intensity.
A direct experiment conducted with ∆Pi = ∆Pc is com-
monly known as being the “unjacketted” compressibil-
ity test. The resulting volumetric strain is then generally
assumed to be the one of the material solid phase i.e. solid
matrix only and the non connected porosity that may be
included inside it.
For a given value of Pc, we proceed as follows:
Step1: An initial confining pressure (P c = 5 MPa)
is applied. Then, porosity experiment is carried out
under the corresponding stress.
Step2: The same gas pressure Pi is applied on both
sides of the sample, which allows the strains to be
measured to evaluate (possible) coupling effects. It
can then be proceeded to the transient pulse decay
method and to evaluate the salt permeability.
Step3: the confining pressure is unloaded Pc from 5
to 0M Pa and strains are measured. Next steps are
then performed at higher P­ c values (Pc = 7, 10, 14,
18, 22 MPa).
3 Experiments
Preliminary remark: as mentioned in the introduction, this
work is the result of almost 18 of months of experiments,
which have also required many steps of new design. Its
objective is nevertheless to take in a wide sweep the main
rock salt properties and not to focus on a specific one. The
results will be used to capture the great tendencies of the salt
behavior. There are, however, major difficulties when testing
such a material with a time dependent behavior and it can be
mentioned that performing quite long permeability measure-
ments on the same sample, in relation with the confining
pressure level, will obviously induce a bias as they cannot
be performed at the same time.
3.1 Gas Permeability Tests
The main purpose of this first series of tests was to evalu-
ate confining pressure effects on gas permeability. As some
coupling effects are suspected (or possible), these tests
were conducted at different levels of mean pore pressure
Pi. The permeability results are, therefore, presented ver-
sus (Pc − Pi), which is the effective Terzaghi stress. Such a
representation enables the detection of couplings upon the
material’s hydraulic behavior. Two samples were used and
for each a whole test was composed of test1 followed by test
2 (see Table 1) allowing some time effects to be detected.
Figure 5 below sumps up the results of gas permeability
measurements under different Pc and Pi for the two sam-
ples and for the two tests (the whole set of results is given

Table 1  Detailed description and experiment schedule of rock salt samples

Rock salt sample Height (mm) Diameter (mm) Weight (g) Test Time

XZ1 61.10 63.80 425.42 Test1 (permeability + porosity + Biot test) 25/06/2018–02/07/2018
Test2 (permeability test) 05/07/2018–20/07/2018
XZ2 52.03 64.43 356.93 Test1 (permeability + porosity test) 28/09/2018–26/10/2018
Test2 (permeability test) 19/11/2018–05/12/2018
Hydrogen and argon permeability test 07/11/2018–19/11/2018
XS1 63.44 64.86 449.18 Poro-mechanical test1(Biot) 23/10/2017–27/10/2017
Poro-mechanical test2(Biot) 14/03/2018–06/04/2018
XS2 63.29 64.50 446.75 Poro-mechanical test (Biot) 29/11/2017–08/12/2017
XZ3 81.80 64.59 594.49 Hydrogen and argon permeability test 11/02/2019–19/02/2019
XS3 81.04 64.28 579.43 Hydrogen and argon permeability test 13/05/2019–16/05/2019
XR3 83.18 64.41 588.04 Triaxial test 09/01/2019–18/01/2019
76.64 66.85 582.52 Hydrostatic test (permeability + Biot test) 08/03/2019–28/03/2019

13
4020 D. Zhang et al.

Sample XZ1 - Test1 Sample XZ1 - Test2

1E-17 1E-18
Pc-Pi (minimum) Pc-Pi (minimum)
Pc-Pi (maximum) Pc-Pi (maximum)

1E-19
Permeability (m )

Permeability (m )
2

2
1E-18

1E-20

1E-19 1E-21
4 6 8 10 12 14 16 0 5 10 15 20 25
Pc (MPa) Pc (MPa)
(a) (b)

Sample XZ2 - Test1 Sample XZ2 - Test2

1E-18
Pc-Pi (minimum) Pc-Pi (minimum)
Pc-Pi(maximum) Pc-Pi(maximum)

1E-17 1E-19
Permeability (m )

Permeability (m )
2

1E-18 1E-20

1E-19 1E-21
0 5 10 15 20 25 0 5 10 15 20 25
Pc (MPa) Pc (MPa)
(c) (d)

Fig. 5  Effect of confining pressure and Terzaghi effective pressure on gas permeability

in “Appendix”). These two samples were extracted from
the same plug XZ and their initial permeability at 3 MPa
of confining pressure is comparable (6.36E−18 m2 and
1.43E−17 m2). A strong effect of confining pressure, lead-
ing to a global reduction of gas permeability by almost three
orders of magnitude, can be observed for both samples at the
end of test 2. The large amount of impermeable crystals in
salt rocks makes the permeability closely related to cracks or
joints. At this scale, the closure of joints around crystals are
likely to play a key role on the flow. As increasing confining
pressure will cause the closure of cracks the latter induces
a strong permeability decrease. The influence of the mean
pore pressure Pi is clearly not systematic. On the first hand,
if present this effect leads to an increase in permeability
when Pi is increased, which excludes a potential Klinken-
berg effect. This means that there is a coupling effect, which
can be attributed to a crack (or joint) aperture effect. On the
second hand, the Pi effect is sometimes present (Fig. 5b, d),

13
Experimental Investigation of Gas Transfer Properties and Stress Coupling Effects of Salt…
partial (Fig. 5a) or negligible (Fig. 5c) and varies with times
as it can be absent for test 1 and present for test 2.
The time effect, which is illustrated in Fig. 6, is clearly
visible as gas permeability, recorded for the test 2 is sys-
tematically lower, by one to two orders of magnitude, than
for test 1. This can be especially underlined at 14 MPa of
confining pressure (in Fig. 6a) and at 22 MPa (in Fig. 6b),
which were the highest confining pressure values reached
in test 1 for the corresponding samples. In fact, the differ-
ences observed with lower confining pressures cannot be
only attributed to the time effect as the hydraulic material’s
behaviour depends on crack closure, which is generally not
reversible. This kind of time effect is consistent with what
is generally reported in the literature (Stormont 1990; Peach
1991; Stormont and Daemen 1992; Maurice et al. 2015).
The time-dependent behavior of salt rock had been demon-
strated through many experiments, which led to constitu-
tive and computational models (Yang et al. 1999; Nazary
Moghadam et al. 2013; Khaledi et al. 2016; Ma et al. 2017;
Mansouri and Ajalloeian 2018; Deng et al. 2020).
X-ray microtomography experiments were conducted on
different salt samples. They allow to observe some identical
parts of samples before and after the various tests. Typical
results are given in Fig. 7 in which the non reversibility of
crack closure can be obviously highlighted.
3.2 Gas Porosity and Confining Pressure
As seen before, gas permeability revealed to be highly sensi-
tive to confining pressure. Porosity measurements (described
Sample XZ1
1E-17
Pc-Pi (maximum)-test1
Pc-Pi (maximum)-test2
Loading Pc
1E-18
Permeability (m )
2
1E-19
Loading Pc
Time effect
1E-20
1E-21
0 5 10 15
Pc (MPa)
(a)
Fig. 6  Comparison of permeability under different Pc between test1 and test2
4021
in Sect. 2.2.2) under confining pressure were, therefore, per-
formed in test1 for the two samples (XZ1 and XZ2) and
the results are given in Fig. 8. The porosity values are in
fact very low since the highest one was 1% at 5 MPa con-
fining pressure. The sensitivity to the confining pressure is
very high as the porosity is almost 0 (or not measurable)
for sample XZ1 at 14 MPa and at 7 MPa for XZ2. Some
interesting deductions can be drawn from such a sensitivity
and relative porosity reduction. The first one is that the salt
cannot be considered as a common porous material. If the
pores were like “classical” ones in usual rocks, the relative
porosity variation would be in the relative order of a few
percent, and not almost one hundred percent as it is here
observed. The second one, which is a consequence of the
first one, is that the (very) small measured pore volumes can
only be constituted of cracks or grain joints between non-
porous crystals. This has been confirmed by SBET analysis
that gave almost no porosity and no specific surface, which
evidenced the fact that there are no pores in the usual sense.
This single porosity measure confirms that fluids can only
flow between the crystals and coupling effects would be
likely take place at these interfaces. Cracking can be then
considered as producing new interfaces.
3.3 Comparison Between Argon and Hydrogen
Permeability
As mentioned in the introduction, salt gas cavern could be
used as hydrogen storage facilities. It is well known that
experiments with hydrogen require safety disposal and
Sample XZ2
1E-16
Pc-Pi(maximum) -test1
Pc-Pi(maximum) -test2
1E-17
Loading Pc
Permeability (m )
2
1E-18
Time effect
1E-19
1E-20
Unloading Pc
1E-21
20 25 0 5 10 15 20 25
Pc (MPa)
(b)
13
4022
Fig. 7  Horizontal tomographic slices of rock salt sample from helical scan (a, c the slices before the hydrostatic test; b, d the corresponding
slices after hydrostatic test)
precautions, which are not always available in rock mechan-
ics laboratories. Leakage of permeability devices is also
among issues that are met with this gas. It is, therefore, very
useful to compare the permeability values obtained with this
gas and a more convenient one like argon. Three tests were
performed on samples having (ultra) low (XZ3), intermedi-
ate (XZ2) and high (XS3) gas permeability. The results are
presented in Fig. 9. The hydrogen permeability was always
the first one to be measured. The process for the sample XZ2
was to confine it at 22 MPa and get the permeability during
the unloading phase. For the two other samples, classical
loading–unloading phases were applied. Test with hydrogen
were performed prior to those with argon. First of all these
results confirm the great sensitivity of salt permeability to
13
D. Zhang et al.
confining pressure and irreversibility, due to crack closure,
which are visible for the unloading paths (Fig. 9b, c). As a
consequence hydrogen and argon permeability comparisons
can be really done at the highest confining pressure. What
can be drawn from these result is that permeability obtained
with both gas are very close to each other and can be consid-
ered as being virtually the same. This interesting result evi-
dences that argon gas can be chosen to evaluate the hydrogen
permeability with no significant differences. This is also true
whatever the order of magnitude of the initial permeability.
The considerable differences obtained on this material (with
permeability being in the range 1­ 0–16–10−23 m2) underline
the material’s dispersion and/or its initial state of cracking.
Experimental Investigation of Gas Transfer Properties and Stress Coupling Effects of Salt…
1,2
Sample XZ1
Sample XZ2
1
0,8
Porosity (%)
0,6
0,4
0,2
0 cracked. On an other hand, the crack closure effect leads to
0 5 10
Pc(MPa)
Fig. 8  Porosity of different salt samples
3.4 Poro‑mechanical Properties (Biot Coefficient
and Coupling Effects)
As mentioned in Sect. 2.2.3, the ratio between the volumet-
ric strain due to an increase in pore pressure with that due
to a confinement unloading can be calculated to evaluate
the ‘apparent’ Biot’s coefficient. In Fig. 10 can be observed
the volumetric strain measured on two samples (XS1 and
XS2), under three different levels of confinement and for a
unit change of 1 MPa of pore pressure. There is clear sam-
ple expansion due to this pore pressure increase, which evi-
dences a poro-mechanical coupling. Increasing the confining
pressure always leads to a reduction of this expansion. As
a consequence, when combined to the volumetric strains
due to a confining pressure decrease (not plotted here), it
was possible to calculate apparent Biot’s coefficients (also
indicated in Fig. 10).
The coefficients obtained are quite low but far from
marginal since they approximately vary from 0.37 to 0.04
under the range of confinement used. The increase, from 7
to 17 MPa, in confining pressure effect results in a reduc-
tion in the Biot’s coefficients, which is currently associated
with a less and less fragmented material, and is likely to
be associated to closure of grain boundaries or cracks. The
time effect is evidenced by comparison of the results given
by two successive identical tests (1 and 2) for the sample
XS1. Hysteretic effects are present at 7 and 12 MPa, between
test 1 and 2, that can be interpreted by the loading history
i.e. 17 MPa were applied during test 1. Such a value might
have led to an irreversible closure of cracks inducing a lower
Biot’s coefficient at 7 and 12 MPa for test 2. On another hand
4023
the coefficients, compared at 17 MPa for both tests, are very
different (0.24 at test 1 and 0.04 at test 2). Such a difference
is due to the time effect and is consistent with what was
observed for the gas permeability (see Fig. 6 for example).
The results of these experiments allow the evaluation of
the bulk modulus Kb and expansion (due to pore pressure
Pi) modulus H. Their variations v.s. confining pressure are
plotted in Fig. 11. They clearly indicate that, whatever is
the confining pressure, the Kb modulus weakly varies in the
range (20–30 GPa) since there is a considerable increase in
the H modulus within the range (50–600 GPa). The confin-
ing pressure is supposed to be linked to crack closure effects.
The results seem to indicate that, for these samples, this
crack closure phenomenon has moderate effects on modulus
Kb. This would mean that these samples are initially weakly
less and less material expansion due to pore pressure. This
15
is consistently linked to the significant increase in the H
modulus.
The latter results were confirmed by other tests on sam-
ple XZ1. Figures 12 and 13 plot the results. The volumetric
strains are given v.s. Terzaghi effective pressure for differ-
ent levels of confining pressure (Fig. 12). At a fixed value
of the latter one can observe the slope εv according Pc − Pi.
A negative slope means that there is expansion due to an
increase in “Pi” while a null (or quasi null) one is relative
to an absence of coupling. The coupling for test 1 is clear
(Fig. 12a) and allowed apparent Biot’s coefficients to be cal-
culated (Fig. 13) whereas it completely vanishes for test 2
(Fig. 12b). This is a new evidence of the combination of irre-
versible crack closure and time effect. It can also be noted
in Fig. 12b (compared with Fig. 12a) that there is a strong
material stiffening as the level of strain εv is lower for test 1
than for test 2. For example it is − 800 μm/m in test 2 and
− 1300 μm/m in test 1 at Pc = 14 MPa (Pi = 0).
4 Triaxial Test (Effects of Damage
on Permeability and Couplings)
As cracks are suspected to lead to significant coupling
effects, a conventional (on a mechanical point of view) tri-
axial test was carried out on sample XR3, then followed by a
series of hydrostatic tests. The main purpose was to damage
the sample with a deviatoric stress and to study the effects of
this damage on gas permeability and couplings with internal
gas pore pressure.
4.1 Permeability Measurement During Triaxial Test
The results presented in Fig. 14 were obtained under
different steps of testing. The initial permeability val-
ues, recorded at step 1 with P c = 3 MPa and at step 2
13
4024 D. Zhang et al.

Sample XZ2 Sample XZ3

1E-18
1E-22
Hydrogen (unloading) Hydrogen (loading)
Argon (unloading) 9E-23
Hydrogen (unloading)
8E-23 Argon (loading)
1E-19 Argon (unloading)
7E-23

6E-23
Permeability (m )
2

1E-20

Permeability(m )
2
5E-23

1E-21
4E-23

1E-22
3E-23

1E-23
0 5 10 15 20 25 0 5 10 15 20 25
Pc(MPa) Pc(MPa)
(a) (b)
Sample XS3
2E-16
Hydrogen (loading)
Hydrogen (unloading)
Argon (loading)
Argon (unloading)
Permeability (m )
2

1E-16

9E-17

8E-17

0 2 4 6 8 10 12
Pc(MPa)
(c)

Fig. 9  Comparison between argon and hydrogen permeability

with Pc = 5 MPa are already extremely low, respectively,
6.1E−23 m 2 and 5.7E−23 m 2, which are virtually the
same. At step 3, the confining pressure was 7 MPa and
the deviatoric loading was started, the permeability fell to
2E-23m2. The following steps were conducted with this
constant confining pressure and the sample degradation
can be observed with the regular increase in permeabil-
ity, by two orders of magnitude, due to the axial stress
increase, up to 44 MPa. Despite this growing, the sample’s
gas permeability remains low: 5E-21m2. The sample was
then dismounted and placed into a hydrostatic cell to per-
formed other tests, described below.
4.2 Permeability and Couplings Measurements
Under Hydrostatic Pressure (After Triaxial Test)
Hydrostatic tests were carried out after triaxial test for the
same sample. Gas permeability was measured and coupling
effects were estimated at 6 levels of confining pressure
(from 3 to 20 MPa). It was also decided to observe the time

13
Experimental Investigation of Gas Transfer Properties and Stress Coupling Effects of Salt… 4025

Strain due to pi - XS2
Strain due to pi - XS1 (test1)
Strain due to pi - XS1 (test2)
25
20
Volumetric strain (10 /MPa)
b - XS2 of confining pressure but it can be underlined that 20 MPa of
b - XS1 (test1) confining pressure are necessary to recover the initial sample
b - XS1 (test2)
permeability i.e. to close the cracks due to deviatoric stress.
0,4 The time effect is still present as it is shown in Fig. 16, which
illustrates the recording of volumetric strains and perme-
ability varying with time, at 14 MPa. There is a first phase
of permeability reduction, which is quite sharp at the begin-
0,3
ning, followed by a plateau. This phenomenon is consistent

with the volumetric strain evolution, which also reached a

-6

Biot coefficient
15
plateau at the end of this phase.
0,2 Results on coupling effects and the resulting apparent
10
Biot’s coefficient are given in Figs. 17 and 18. As for the
previous experiments, the potential coupling effect was
0,1
detected by a progressive increase in gas pore pressure,
5 while the confining pressure is fixed at a constant value. At
the highest confining pressure (20 MPa) almost no coupling
was observed i.e. the volumetric strain due to Pi is virtually
0 0 0. For the other confining pressures (4–14 MPa), there is
5 10 15 20
Pc(MPa) a clear coupling which is illustrated both in Figs. 17 and
18. The effect of Terzaghi effective pressure is also present
Fig. 10  Biot coefficient and volumetric strain due to Pi (Fig. 17) as there is a stronger coupling for the lowest val-
ues (i.e. high Pi level) than for the highest ones. There is an
apparent threshold of 7 MPa, which is only indicative here.
The volumetric strains due to Pi and the apparent Biot’s coef-
ficient can be observed in Fig. 18. The confining pressure
level and the crack closure induced have a strong reduction
effect of the volumetric expansion, which exhibits a continu-
ous decrease. This kind of phenomenon was also observed
on cracked sandstone samples (Hu et al. 2018). Combined
with the progressive material stiffening, this leads to a stable
value on the apparent Biot’s coefficient, lying in the range
0.35–0.3 (excepted at 5 MPa, which is maybe artifact). This
Biot’s coefficient value is significantly higher than for the
previous experiments, especially at medium confining pres-
sure (7–14 MPa).

5 Conclusion

In this study, salt gas permeability was measured under

hydrostatic or triaxial loadings. Coupling effects between
internal fluid pressure and volumetric strains have been evi-
denced. These results make a great advance in the under-
Fig. 11  Variations of moduli H and Kb due to confining pressure
standing of salt behaviour as they were generally considered
as non-existent or negligible for this kind of rock. Salt is a
effect at Pc = 14 MPa. The results for gas permeability can complex material and very difficult to characterize. Not only
be found in Fig. 15 and the effect of time in Fig. 16. At the permeability is very low and sensitive to confinement but
Pc = 3 MPa the permeability is now 2.8E−19 m2. This value it also evolves with time because of the viscous effects, even
can be compared to the first one, measured during the triax- for isotropic loading. Beyond that, permeabilities measured
ial test, which was around 6.1E-23m2. Such an increase, by 4 with argon or hydrogen are close to each other, and argon
orders of magnitude, indicates that the sample was seriously permeability may be a good estimate of hydrogen permeabil-
damaged and that multiple cracks had been produced. There ity in studies for hydrogen storage development. This result
is then a classical decrease in permeability with the increase will allow easier measurements of gas permeability that will

13
4026 D. Zhang et al.

Sample XZ1 (Test1) Sample XZ1 (Test2)

-200 -200
Pc=5MPa Pc=5MPa
Pc=7MPa Pc=7MPa
Pc=10MPa Pc=10MPa
-400 Y=-355-31.5X Pc=14MPa Pc=14MPa
-400
Pc=18MPa
Pc=22MPa

-600
Volumetric strain (10 )

Y=-502.69-26.812X

Volumetric strain (10 )

-6

-6
-600

-800
-800
Y=-802.73-19.388X

-1000
-1000

-1200
Y=-1164.3-10.393X
-1200

-1400
0 5 10 15 0 5 10 15 20 25
Pc-Pi (MPa) Pc-Pi (MPa)
(a) (b)

Fig. 12  Variation of volumetric strain with (Pc − Pi)

Sample XZ1 (Test1) 1E-20 50

40 0,4
Permeability (m2)
Volumetric Strain due to Pi Biot coefficient
Confining Pressure (MPa)

35 deviatoric stress (MPa) 40

1E-21
Volumetric Strain (10 /MPa)

Permeability (m )
2

30 0,3

Stress (MPa)
30
-6

Biot coefficient

25
20
1E-22
20 0,2
10

15
1E-23 0
0 1 2 3 4 5 6 7 8
10 0,1 Test steps
4 8 12 16
Pc (MPa)
Fig. 14  Permeability measurement during triaxial test
Fig. 13  Biot coefficient and volumetric strain due to Pi
permeable sandstones, shales or other sedimentary rocks.
They are also sensitive to the material’s damage obtained
be involved in the hydrogen storage concept. Concerning under triaxial test and the first tendency is an increase in the
pore pressure coupling effects, it has been experimentally latter due to crack generation.
verified that salt rock displays a poromechanical behavior. At a microscopic scale, the sensitivity to confinement
Fluid pressure can induce rock skeleton deformation and of the permeability and the Biot’s coefficient support the
will have to take place in the stress–strain equation of state hypothesis that fluid flows through cracks or boundaries
characterizing rock salt and the permeability. The estimated between salt grains. Further experiments with joint study
Biot’s coefficients are sensitive to confinement but they on microstructure are planned to verify this hypothesis. The
are far below usual values measured for permeable or low future of this study will also concern a precise measurement

13
Experimental Investigation of Gas Transfer Properties and Stress Coupling Effects of Salt… 4027

1E-18 Pc =3MPa
0
Pc = 5MPa
Permeability Pc =7MPa
Pc =10MPa
Y= -257,07 - 47,457X Pc =14MPa
1E-19 -500 Pc =14MPa (after204hr )
Pc =20 (MPa)

Y= -582,13 - 38,95X

Volumetric strain (10 )

-6
Permeability (m )

1E-20 -1000
2

Y=-763,86 - 37,413X

Y=-1045,6 - 21,244X

1E-21 -1500
Y=-1360,3 - 22,511X

Y= -1571 - 19,64X
Time Interval = 204hours
-2000
1E-22

-2500
1E-23 0 5 10 15 20 25
0 5 10 15 20 25 Pc - Pi (MPa)
Pc (MPa)

Fig. 17  Variation of volumetric strain with (Pc − Pi)

Fig. 15  Variation of permeability with Pc

Volumetric Strain due to Pi Biot coefficient

Permeability Volumetric strain 50 0,5
8E-22 -1600

-1632 40
6E-22
Volumetric Strain (10 /MPa)

Pc = 14MPa 0,4
-6
Volumetric strain (10 )

Biot coefficient
Permeability (m )

-1664
2

4E-22 30

-1696
0,3
-6

2E-22 20
-1728

0 -1760
0 50 100 150 200 10 0,2
Time (hr) 0 4 8 12 16
Pc (MPa)

Fig. 16  Variation of permeability and strain with time

Fig. 18  Apparent Biot’s coefficient

of permeability under deviatoric loading with a particular increase in brine pressure will bring about poromechanical
attention paid to the influence of creep and the potential effects, that may affect the permeability around the cavern.
sealing of cracks with time.
However, it is obvious that in-situ tests concerning the
abandonment of caverns take place on a very different scale Funding This work is related to Dongmei Zhang PhD and financially
supported by the Chinese Scholarship Council. The authors wish to
from that of laboratory samples. This experimental study thank Storengy for its financial support and for the salt core supplying.
nevertheless shows that gas permeability in the cavern vicin-
ity is likely to increase by several orders of magnitude higher Compliance with Ethical Standards
than the virgin permeability of the salt rock mass. Cavern
development indeed induces deviatoric stress able to pro- Conflict of interest No potential conflict of interests was reported by
duce cracks around the cavern. During abandonment, the the authors.

13
4028 D. Zhang et al.

Appendix Berest P, Brouard B, Durup JG (2001) Tightness tests in salt-cavern

wells. Oil Gas Sci Technol 56:451–469. https​://doi.org/10.2516/
ogst:20010​37
The complete results of gas permeability measurements of Biot MA (1941) General theory of three-dimensional consolidation. J
two tests (test 1 and test 2) for samples XZ1 and XZ2 are Appl Phys 12:155–164. https​://doi.org/10.1063/1.17128​86
shown in Table 2. Brodsky NS (1993) Hydrostatic and shear consolidation tests with per-
meability measurements on Waste Isolation Pilot Plant crushed
salt. SAND93-7058, Sandia National Laboratories
Chan KS, Bodner SR, Munson DE (2001) Permeability of WIPP salt
Table 2  Gas permeability measurements under different P
­ c and P
­ i for during damage evolution and healing. Int J Damage Mech 10:347–
two samples 375. https​://doi.org/10.1106/H3UV-1URA-AFUY-FX49
Chen J, Ren S, Yang C et al (2013a) Self-healing characteristics of
Pc (MPa) Pi (MPa) K ­(m2)—sample XZ1 K ­(m2)—sample XZ2 damaged rock salt under different healing conditions. Materials
6:3438–3450. https​://doi.org/10.3390/ma608​3438
Test1 Test2 Test1 Test2 Chen X, Caratini G, Davy CA et al (2013b) Coupled transport and
poro-mechanical properties of a heat-treated mortar under con-
5 1 4.17E−18 4.13E−19 1.34E−17 4.61E−20
finement. Cem Concr Res 49:10–20. https​://doi.org/10.1016/j.
5 3 3.34E−18 4.42E−19 1.40E−17 8.15E−20 cemco​nres.2013.03.006
5 4 1.41E−17 1.39E−19 Deng J, Liu Y, Yang Q et al (2020) A viscoelastic, viscoplastic, and
7 2 1.62E−18 1.44E−19 1.24E−17 4.62E−20 viscodamage constitutive model of salt rock for underground
energy storage cavern. Comput Geotech 119:103288. https​://doi.
7 4 1.70E−18 2.58E−19 1.14E−17 6.97E−20
org/10.1016/j.compg​eo.2019.10328​8
7 5 1.43E−17 1.13E−19 Fedor F, Hámos G, Jobbik A et al (2008) Laboratory pressure pulse
10 2 1.43E−18 6.41E−20 9.73E−18 4.66E−20 decay permeability measurement of Boda Claystone, Mecsek
10 4 1.70E−18 7.46E−20 7.19E−18 6.94E−20 Mts., SW Hungary. Phys Chem Earth Parts A/B/C 33:S45–S53.
https​://doi.org/10.1016/j.pce.2008.10.059
10 6 1.74E−18 1.14E−19 9.40E−18 9.42E−20
Frédéric S, Olivier C, Zouber L (2003) Sur la fiabilité des mesures des
10 8 1.75E−18 2.17E−19 1.08E−17 1.08E−19 perméabilités hétérogènes par injection de gaz. Micro-pulse tests.
14 2 1.69E−19 5.81E−21 4.16E−18 1.27E−20 Rev Franç Génie Civ 7:451–469
14 4 1.70E−19 8.14E−21 4.28E−18 2.08E−20 Gloyna EF, Reynolds TD (1961) Permeability measurements of rock
salt. J Geophys Res 66:3913–3921. https:​ //doi.org/10.1029/JZ066​
14 6 2.31E−19 1.12E−20 4.79E−18 4.27E−20
i011p​03913​
14 8 3.05E−19 1.26E−20 3.66E−18 3.60E−20 Hou Z (2003) Mechanical and hydraulic behavior of rock salt in the
14 10 3.50E−19 1.72E−20 5.81E−18 5.77E−20 excavation disturbed zone around underground facilities. Int J
14 12 5.86E−19 2.87E−20 6.34E−18 7.22E−20 Rock Mech Min Sci 40:725–738. https​://doi.org/10.1016/S1365​
-1609(03)00064​-9
18 2 4.35E−21 2.23E−18 1.14E−20
Hsieh PA, Tracy JV, Neuzil CE et al (1981) A transient laboratory
18 4 6.39E−21 2.20E−18 2.76E−20 method for determining the hydraulic properties of “tight” rocks—
18 6 5.51E−21 3.41E−18 1.89E−20 I. Theory. Int J Rock Mech Min Sci Geomech Abstr 18:245–252
18 8 8.49E−21 3.27E−18 2.15E−20 Hu C, Agostini F, Skoczylas F et al (2018) Poromechanical properties
of a sandstone under different stress states. Rock Mech Rock Eng
18 10 1.03E−20 2.93E−18 2.90E−20
51:3699–3717. https​://doi.org/10.1007/s0060​3-018-1550-x
18 12 1.15E−20 2.45E−18 4.83E−20 Karimi-Jafari M. JM (2007) Sur le comportement transitoire des cavités
22 2 2.88E−21 8.30E−19 3.70E−21 salines profondes. Dissertation, Ecole polytechnique
22 4 4.78E−21 8.13E−19 3.77E−21 Khaledi K, Mahmoudi E, Datcheva M et al (2016) Sensitivity analy-
sis and parameter identification of a time dependent constitutive
22 6 4.41E−21 8.57E−19 1.90E−20
model for rock salt. J Comput Appl Math 293:128–138. https​://
22 8 4.45E−21 8.22E−19 2.17E−20 doi.org/10.1016/j.cam.2015.03.049
22 10 6.82E−21 8.67E−19 2.03E−20 Labaune P, Rouabhi A, Tijani M et al (2018) Dilatancy criteria for
22 12 1.16E−20 9.72E−19 2.19E−20 salt cavern design: a comparison between stress- and strain-
based approaches. Rock Mech Rock Eng 51:599–611. https​://doi.
org/10.1007/s0060​3-017-1338-4
Liang Y, Price JD, Wark DA, Watson EB (2001) Nonlinear pressure
diffusion in a porous medium: approximate solutions with applica-
References tions to permeability measurements using transient pulse decay
method. J Geophys Res Solid Earth 106:529–535. https​://doi.
Alkan H (2009) Percolation model for dilatancy-induced permeability org/10.1029/2000J​B9003​44
of the excavation damaged zone in rock salt. Int J Rock Mech Min Liu W, Li Y, Yang C et al (2015) Permeability characteristics of mud-
Sci 46:716–724. https​://doi.org/10.1016/j.ijrmm​s.2008.08.002 stone cap rock and interlayers in bedded salt formations and tight-
Bérest P, Brouard B (2003) Safety of salt caverns used for under- ness assessment for underground gas storage caverns. Eng Geol
ground storage blow out; mechanical instability; seepage; cav- 193:212–223. https​://doi.org/10.1016/j.engge​o.2015.04.010
ern abandonment. Oil Gas Sci Technol 58:361–384. https​://doi. Ma L, Wang M, Zhang N et al (2017) A variable-parameter creep
org/10.2516/ogst:20030​23 damage model incorporating the effects of loading frequency for
Bérest P, Bergues J, Brouard B et al (2001) A salt cavern abandon- rock salt and its application in a bedded storage cavern. Rock
ment test. Int J Rock Mech Min Sci 38:357–368. https​://doi. Mech Rock Eng 50:2495–2509. https​://doi.org/10.1007/s0060​
org/10.1016/S1365​-1609(01)00004​-1 3-017-1236-9

13
Experimental Investigation of Gas Transfer Properties and Stress Coupling Effects of Salt…
Mansouri H, Ajalloeian R (2018) Mechanical behavior of salt rock
under uniaxial compression and creep tests. Int J Rock Mech Min
Sci 110:19–27. https​://doi.org/10.1016/j.ijrmm​s.2018.07.006
Maurice S, Bannach A, Amro M (2015) Renewable energy storage in
salt caverns—a comparison of thermodynamics and permeability
between natural gas, air and hydrogen. Solution Mining Research
Institute, Germany
Nazary Moghadam S, Mirzabozorg H, Noorzad A (2013) Modeling
time-dependent behavior of gas caverns in rock salt considering
creep, dilatancy and failure. Tunn Undergr Sp Technol 33:171–
185. https​://doi.org/10.1016/j.tust.2012.10.001
Peach CJ (1991) Influence of deformation on the fluid transport proper-
ties of salt rocks. Dissertation, University of Utrecht
Peach JC, Speirs CJ, Tankink AJ, Zwart HJ (1987) Fluid and ionic
transport properties of deformed salt rock. Commission of the
European Communities Report EUR10926, 277–290
Popp T, Kern H (1998) Ultrasonic wave velocities, gas permeability
and porosity in natural and granular rock salt. Phys Chem Earth
23:373–378. https​://doi.org/10.1016/S0079​-1946(98)00040​-8
Pusch G, Clausthal T, Alkan H (2002) Gas permeation models related
to dilatancy development under deviatoric stress conditions. Solu-
tion Mining Research Institute, Fall 2002 Technical Meeting Bad
Ischl, Austria, October 6–9, p 18
Schulze O, Popp T, Kern H (2001) Development of damage and perme-
ability in deforming rock salt. Eng Geol 61:163–180
Stolten D, Grube T (eds) (2010) 18th world hydrogen energy confer-
ence 2010—WHEC 2010: proceedings. 4: Parallel sessions book:
storage systems, policy perspectives, initiatives and cooperations.
Forschungszentrum, IEF-3, Jülich
Stormont JC (1990) Gas permeability changes in rock salt during defor-
mation. Dissertation, The University of Arizona
Stormont JC (1997) Conduct and interpretation of gas permeability
measurements in rock salt. Int J Rock Mech Min Sci 34:3–4
4029
Stormont JC, Daemen JJK (1992) Laboratory study of gas permeability
changes in rock salt during deformation. Int J Rock Mech Min
Sci 29:325–342
Terzaghi K (1943) Theoretical soil mechanics. Wiley, New York
Urai JL, Schléder Z, Spiers CJ, Kukla PA (2008) Flow and transport
properties of salt rocks. In: Littke R (ed) Dynamics of complex
intracontinental basins: the Central European Basin System.
Springer, Berlin
Walder J, Nur A (1986) Permeability measurement by the pulse-decay
method: effects of poroelastic phenomena and non-linear pore
pressure diffusion. Int J Rock Mech Min Sci Geomech Abstr
23:225–232. https​://doi.org/10.1016/0148-9062(86)90968​-X
Wang Y, Jeannin L, Agostini F et al (2018) Experimental study and
micromechanical interpretation of the poroelastic behaviour
and permeability of a tight sandstone. Int J Rock Mech Min Sci
103:89–95. https​://doi.org/10.1016/j.ijrmm​s.2018.01.007
Xing W, Zhao J, Düsterloh U et al (2014) Experimental study of
mechanical and hydraulic properties of bedded rock salt from the
Jintan location. Acta Geotech 9:145–151
Yang C, Daemen JJK, Yin J-H (1999) Experimental investigation of
creep behavior of salt rock. Int J Rock Mech Min Sci 36:233–242.
https​://doi.org/10.1016/S0148​-9062(98)00187​-9
Yuan H, Agostini F, Duan Z et al (2017) Measurement of Biot’s coef-
ficient for COx argillite using gas pressure technique. Int J Rock
Mech Min Sci 92:72–80
Publisher’s Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
13